Communication between binding sites is required for YqjI regulation of target promoters within the yqjH-yqjI intergenic region

Abstract

The nickel-responsive transcription factor YqjI represses its own transcription and transcription of the divergent yqjH gene, which encodes a novel ferric siderophore reductase. The intergenic region between the two promoters is complex, with multiple sequence features that may impact YqjI-dependent regulation of its two target promoters. We utilized mutagenesis and DNase I footprinting to characterize YqjI regulation of the yqjH-yqjI intergenic region. The results show that YqjI binding results in an extended footprint at the yqjI promoter (site II) compared to the yqjH promoter (site I). Mutagenesis of in vivo gene reporter constructs revealed that the two YqjI binding sites, while separated by nearly 200 bp, appear to communicate in order to provide full YqjI-dependent regulation at the two target promoters. Thus, YqjI binding at both promoters is required for full repression of either promoter, suggesting that the two YqjI binding sites cooperate to control transcription from the divergent promoters. Furthermore, internal deletions that shorten the total length of the intergenic region disrupt the ability of YqjI to regulate the yqjH promoter. Finally, mutagenesis of the repetitive extragenic palindromic (REP) elements within the yqjH-yqjI intergenic region shows that these sequences are not required for YqjI regulation. These studies provide a complex picture of novel YqjI transcriptional regulation within the yqjH-yqjI intergenic region and suggest a possible model for communication between the YqjI binding sites at each target promoter.

FIG 1

Summary of regulatory features within the yqjH-yqjI intergenic region. The nucleotide sequence numbering begins with the first nucleotide upstream of the yqjH start codon and ends with the first nucleotide prior to the yqjI start codon. Regions of the intergenic sequence protected from DNase I cleavage by apo-YqjI are indicated in black boxes (see the text and Fig. 2). Light gray boxes with dashed outlines indicate the three potential Fur binding sites proposed from bioinformatic analysis (9, 10). The originally identified palindromic elements are indicated by arrows within the YqjI footprints (5). REP228a, and REP228b elements are indicated with bracketed lines. The −35 and −10 RNA polymerase binding sites are in boldface font, and the +1 transcriptional start sites are in boldface italic font. The +1 transcriptional start site for the yqjH promoter has not been experimentally determined and is only predicted.

FIG 2

DNase I footprinting of apo-YqjI binding to the yqjH (A) and yqjI (B) promoters on the full-length yqjH-yqjI intergenic region. The black traces are the elution profiles of DNase I digests of 2 nM labeled DNA from the noncoding strand of the promoter. Red traces are the elution profiles of DNase I digests of 2 nM labeled promoter DNA after incubation with 200 nM oligomeric apo-YqjI. The numbering along the top of each panel indicates the nucleotide positions of cleavage sites (that generate the observed fragments) on the noncoding strand relative to the transcriptional start site (+1) of each promoter. Underlined nucleotide sequences below each panel indicate the apo-YqjI binding sites at each promoter and are specific for the noncoding strand. The dotted underline indicates the extended footprint of site II relative to the original palindrome. RFU, relative fluorescence units. (C) Duplex DNA of sites I and II aligned to show the similarities between the site I half-site and the downstream half-site of site II. Boldface red font indicates nucleotides protected by YqjI in the DNase I footprints in panels A and B. Since sites I and II are divergent in the genome, sequences are oriented with the noncoding strand of each promoter on top to facilitate alignment.

FIG 3

Effect of mutations in site I on YqjI repression of yqjH-lacZ fusions in vivo. The top row is a diagram of the wild-type construct containing the full-length intergenic region with the yqjH-lacZ fusion. Site I and II operators are shown as open arrows. Specific nucleotide changes made to site I are shown in each of the lower rows on the coding strand of the promoter region. All mutations were made on the full-length construct shown in the top row. Boldface font indicates nucleotides protected by YqjI in DNase I footprinting (see Fig. 2). All constructs were single-copy reporters in wild-type (MG1655) E. coli. Promoter activity is indicated as Miller units, and each value is the average from triplicate experiments.

FIG 4

Quantitative DNase I footprinting for each YqjI binding site. (A) K_d values were calculated based on DNase I protection upon oligomeric apo-YqjI binding at site I or site II operators when using the partial or full-length intergenic region for footprinting. Half-sites within the site II operator were characterized separately, with “down” indicating the downstream half-site and “up” indicating the upstream half-site relative to the direction of yqjI transcription. (B) The sums of peak heights for select protected peaks in each site were plotted as a function of the apo-YqjI concentration for the entire apo-YqjI titration data set (filled circles). Solid lines indicate nonlinear regression fit using the sigmoidal dose-response model provided by GraphPad Prism.

FIG 5

Effect of mutations in distal YqjI binding sites on YqjI repression of proximal promoter-lacZ fusions in vivo. (A) Dashed gray outlines of arrows indicate deletion of the half-site(s) within the site II operator (open arrows); the spacer regions between the half-sites are shown as dashed gray lines. (B) Asterisks indicate mutations within the site I operator (with the half-sites shown as open arrows) as described in Fig. 3 and Fig. S2 in the supplemental material. All constructs were single-copy reporters in the wild-type (MG1655) or ΔyqjI strain of E. coli. Promoter activity is indicated as Miller units, and each value is the average from triplicate experiments.

FIG 6

Effect of REP element mutations on YqjI repression of yqjH-lacZ (A) or yqjI-lacZ (B) fusions in vivo. Dashed gray lines indicate deletion of REP 228b (which shortens the overall length of the intergenic region). Asterisks indicate mutations that disrupt the palindromes within the indicated REP element(s) but do not change the length of the intergenic region (see Fig. S1 for specific nucleotide changes). Site I and II operators are shown as open arrows. All constructs were single-copy reporters in wild-type (MG1655) E. coli. Promoter activity is indicated as Miller units, and each value is the average from triplicate experiments.